A fast analysis method for satellite downlink data
By employing multi-core parallel processing of CPU and a producer-consumer model, combined with image gradient changes, the problem of low data parsing efficiency in satellite downlink data transmission has been solved, achieving fast and efficient data parsing suitable for high-resolution solar observation data processing.
Patent Information
- Application Number
- CN202411231787.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-09-04
AI Technical Summary
Existing satellite downlink data parsing methods are inefficient, especially with large data volumes. Traditional CPU multithreading and GPU parallel computing suffer from copy overhead and frequent I/O operations, resulting in excessively long parsing times.
The system employs CPU-based multi-core parallel processing, combined with a producer-consumer model, to separate data parsing from disk read/write operations. It determines the scanning sequence through image gradient changes, uses multi-process parallel processing and the JPEG2000 algorithm to parse spectral data, and generates a full-surface image.
It improves the resolution speed of satellite downlink data, reduces computational complexity, and is robust and effective, making it suitable for processing solar observation data with high temporal and spatial resolution.
Smart Images

Figure CN118761940B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of astronomy technology and image processing technology, and in particular to a method for quickly parsing satellite downlink data. Background Art
[0002] The sun is the only star that can be observed with high temporal and spatial resolution. Through observation and research of the sun, exploring methods and approaches for space weather forecasting has important practical significance for national defense and the national economy. Compared with ground-based solar telescopes, solar space exploration has the advantages of not being affected by the turbulence and absorption effects of the earth's atmosphere and can be observed continuously. Especially driven by science and technology, solar exploration satellites have developed rapidly.
[0003] Since solar observations require high temporal resolution, high spatial resolution, and high spectral resolution, satellite payloads will collect and generate a large amount of data. After being encoded by the onboard data compression unit, these data are temporarily stored in the onboard storage unit and are waiting to be transmitted to the ground application system when the satellite passes by. This requires that on the one hand, the data must be received as quickly as possible, and on the other hand, the scientific application system must first quickly decode these raw data after receiving them, so that subsequent scientific data calibration can be carried out for use by scientific researchers and subsequent data analysis.
[0004] At present, the common way to parse data is to read data sequentially. This method is relatively simple to implement, and the staff can quickly develop programs. This method mainly runs on the Windows platform and focuses on verifying the performance indicators of the payload during the on-orbit test. In this way, if a data packet of one million frames is parsed, it often takes dozens of hours. In addition, there are multi-threaded and multi-process parallel processing methods based on the CPU, but generally require accurate control of the starting position of the image. First, divide the image into blocks and decode each block. This often requires scanning the file once to determine the starting position of the image, which is very time-consuming.
[0005] In recent years, with the development of GPU, GPU has many CUDA computing cores, which have significant advantages in large-scale parallel computing tasks. It has also begun to be used in decoding work, especially in image reconstruction and processing. GPU can easily decompose large-scale matrix calculations into small vector calculations and directly parallelize them, which significantly improves the speed. However, for large amounts of data, especially logically related data calculations, GPU also has a data copy overhead from memory to video memory. Considering the copy overhead, it will take more time. At the same time, considering the frequent IO operations, this will also reduce efficiency.
[0006] Considering the massive amount of data, it is very necessary to be able to quickly complete data analysis. The present invention is to solve this problem and proposes multi-core parallelism based on CPU, adopts the producer-consumer positive model to ensure process synchronization, and separates IO operations and CPU calculations to achieve fast data analysis. This provides a universal data processing method for space solar exploration or other large data volume tasks. Summary of the invention
[0007] The purpose of the present invention is to provide a method for quickly parsing satellite downlink data.
[0008] The object of the present invention can be achieved by the following technical scheme: A method for quickly parsing satellite downlink data, comprising the following steps:
[0009] Step S1: receiving and decompressing the raw data to be processed to obtain decompressed data, segmenting the obtained decompressed data to obtain segmented data, extracting data from each segmented data to obtain spectral data, and analyzing and synthesizing the extracted spectral data to obtain a spectral image;
[0010] Step S2: obtaining a corresponding image intensity sequence according to the analyzed spectral image, performing gradient calculation on the obtained image intensity sequence to obtain a corresponding image gradient sequence, searching for a maximum image frame number in the obtained image gradient sequence, obtaining a scanning sequence according to the searched maximum image frame number, and then generating a full disk image;
[0011] Step S3: Perform a traversal operation on the generated full disk image to obtain full disk pixel values, adopt a binary search method to calculate the obtained full disk pixel values to obtain left and right extreme points and upper and lower extreme points, calculate the candidate heliocentric pixel positions based on the obtained left and right extreme points and upper and lower extreme points, and record the corresponding number of occurrences, obtain the heliocentric pixel coordinates based on the recorded number of occurrences, obtain the geometric distance between the heliocentric pixel coordinates and the image center coordinates, and adjust the generated full disk image based on the obtained geometric distance.
[0012] Further, the process of receiving and decompressing the original data to be processed to obtain decompressed data, and segmenting the obtained decompressed data to obtain segmented data includes:
[0013] The original data transmitted by the satellite is received, and the Tar command provided by the Linux system is used to decompress the received original data to obtain corresponding decompressed data, and the obtained decompressed data is segmented, wherein the segmentation refers to dividing the decompressed data into a plurality of segmented data.
[0014] Furthermore, the process of extracting data from each segmented data to obtain spectral data, and analyzing and synthesizing the extracted spectral data to obtain a spectral image includes:
[0015] A multi-process parallel processing method is used to extract data from each segmented data. The data extraction refers to extracting the spectral data from the original data. One segmented data corresponds to one data extraction process, and the position of the scientific payload data in the segmented data is obtained. The spectral data and auxiliary information therein are extracted. The JPEG2000 algorithm is used to analyze the six segments of spectral data respectively, and the six segments of spectral images that have been analyzed are synthesized to obtain a complete spectral image.
[0016] Create a data extraction process pool to manage the data extraction process, create a buffer to hold the parsed spectral image, and use the producer-consumer model to manage the data extraction process and buffer. The producer process is responsible for parsing the image, and the consumer process is responsible for reading and writing data to the disk.
[0017] Furthermore, a process of obtaining a corresponding image intensity sequence according to the analyzed spectral image and performing gradient calculation on the obtained image intensity sequence to obtain a corresponding image gradient sequence includes:
[0018] The image intensity of each frame of the obtained spectral image is calculated to obtain the corresponding image intensity sequence, which is recorded as {I0, I1, ..., I n-1};
[0019]
[0020] Where w is the width of the spectral image, h is the height of the spectral image, for The intensity value of the corresponding pixel at the position is calculated, and the gradient of the obtained image intensity sequence is calculated to obtain the corresponding image gradient sequence, which is recorded as {T0, T1, ..., T n-1};
[0021]
[0022] Furthermore, the process of searching for the maximum image frame number in the obtained image gradient sequence, obtaining a scanning sequence according to the found maximum image frame number, and then generating a full disk image includes:
[0023] In the image gradient sequence, the image frame number corresponding to the position with the largest gradient change is taken as the maximum image frame number. The maximum image frame number includes two. A scanning sequence is determined according to the two maximum image frame numbers obtained. For a scanning sequence, it is spliced according to its internal wavelength order to obtain the corresponding full-disk image.
[0024] Furthermore, the generated full disk image is traversed to obtain full disk pixel values, and the process of using a binary method to calculate the obtained full disk pixel values to obtain left and right extreme value points and upper and lower extreme value points includes:
[0025] A traversal operation is performed on the obtained full disk image, and the traversal operation is used to visit each pixel point in the full disk image one by one to obtain the corresponding full disk pixel value, and the obtained full disk pixel value is calculated by using the binary method, the number of pixels in each row is divided equally, and the calculation is started from the starting and ending positions of each row to obtain the corresponding left and right extreme value points, and the upper and lower extreme value points are obtained by the same method.
[0026] Further, the candidate heliocentric pixel positions are calculated based on the obtained left and right extreme value points and the upper and lower extreme value points, and the corresponding number of occurrences is recorded. The process of obtaining the heliocentric pixel coordinates based on the recorded number of occurrences includes:
[0027] The coordinates of the left and right extreme points of each group are obtained, and the candidate heliocentric pixel positions are obtained according to the coordinates of the left and right extreme points, and the number of occurrences of each candidate heliocentric pixel position is recorded; the coordinates of the upper and lower extreme points of each group are obtained, and the candidate heliocentric pixel positions are obtained according to the coordinates of the upper and lower extreme points, and the number of occurrences of each candidate heliocentric pixel position is recorded; the candidate heliocentric pixel position with the largest number of occurrences is used as the heliocentric pixel position of the full-sun disk image, and the corresponding heliocentric pixel coordinates are output.
[0028] Furthermore, the geometric distance between the heliocentric pixel coordinates and the image center coordinates is obtained, and the process of adjusting the generated full disk image according to the obtained geometric distance includes:
[0029] The image center of the full-disk image and its corresponding image center coordinates are obtained, the geometric distance between the two is obtained according to the obtained heliocentric pixel coordinates and the image center coordinates, the azimuth information of the heliocentric pixel coordinates relative to the image center coordinates is obtained, the azimuth information is used to describe the deviation angle of the heliocentric pixel coordinates relative to the image center coordinates, and the obtained full-disk image is moved by a corresponding geometric distance according to the deviation angle to achieve adjustment.
[0030] Compared with the prior art, the present invention has the following beneficial effects:
[0031] 1. By separating data analysis from disk read and write operations and adopting a multi-core parallel mechanism, the speed of raw data analysis is improved. Based on the characteristics of solar observation images, a scanning sequence is determined by utilizing the characteristics of image gradient changes. The algorithm has good robustness and is insensitive to solar activity. The gradient method is used to determine the center of the solar disk. Compared with traditional image cross-correlation and Hough circle fitting methods, the computational complexity is low and the processing speed is improved. This method can be used in circle center search and full-disk guidance of the telescope. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION
[0033] like Figure 1 As shown, a method for quickly parsing satellite downlink data includes the following steps:
[0034] Step S1: receiving and decompressing the raw data to be processed to obtain decompressed data, segmenting the obtained decompressed data to obtain segmented data, extracting data from each segmented data to obtain spectral data, and analyzing and synthesizing the extracted spectral data to obtain a spectral image;
[0035] Step S2: obtaining a corresponding image intensity sequence according to the analyzed spectral image, performing gradient calculation on the obtained image intensity sequence to obtain a corresponding image gradient sequence, searching for a maximum image frame number in the obtained image gradient sequence, obtaining a scanning sequence according to the searched maximum image frame number, and then generating a full disk image;
[0036] Step S3: Perform a traversal operation on the generated full disk image to obtain full disk pixel values, adopt a binary search method to calculate the obtained full disk pixel values to obtain left and right extreme points and upper and lower extreme points, calculate the candidate heliocentric pixel positions based on the obtained left and right extreme points and upper and lower extreme points, and record the corresponding number of occurrences, obtain the heliocentric pixel coordinates based on the recorded number of occurrences, obtain the geometric distance between the heliocentric pixel coordinates and the image center coordinates, and adjust the generated full disk image based on the obtained geometric distance.
[0037] It should be further explained that, in a specific implementation process, the process of receiving and decompressing the original data to be processed to obtain decompressed data, and segmenting the obtained decompressed data to obtain segmented data includes:
[0038] The original data transmitted by the satellite is received. At this time, the original data is in a state to be processed. The Tar command provided by the Linux system is used to decompress the received original data to obtain the corresponding decompressed data;
[0039] Under normal circumstances, it takes 4624 steps for a satellite to perform an observation scan, and it takes 46 seconds to complete an observation scan. The spectral data collected in each frame is divided into 6 segments. Each segment of spectral data must be divided into several data frames for storage and download using a preset image data format. Each raw data received by the ground application system usually contains hundreds of thousands to millions of frames of spectral data.
[0040] The image data format includes "synchronization header", "leading header", "scientific payload data", and "error control", wherein the synchronization header is 4 bytes, the leading header is 8 bytes, the scientific payload data is 2032 bytes, and the error control is 4 bytes;
[0041] The obtained decompressed data is segmented, and the segmentation processing refers to dividing the decompressed data into several sub-data based on the segmentation standard of 10MB. If the file size of the decompressed data is 1.98GB, and the file size of each sub-data is 10MB, then the decompressed data will be divided into approximately 2000 sub-data, among which the file size of the last sub-data is less than or equal to 10MB, and the divided sub-data will be marked as segmented data.
[0042] It should be further explained that, in the specific implementation process, the process of extracting data from each segmented data to obtain spectral data, and parsing and synthesizing the extracted spectral data to obtain a spectral image includes:
[0043] A multi-process parallel processing method is used to extract data from each segmented data. The data extraction refers to extracting the spectral data from the original data. One segmented data corresponds to one data extraction process. In the embodiment of the present invention, nearly 2000 data extraction processes are required. Each data extraction process will search for the first frame header, i.e., the synchronization header, in the segmented data corresponding to itself to ensure that there will be no data loss in the spectral data across the segmented data. In this way, each data extraction process can start from the starting position of a complete frame of spectral data.
[0044] Taking the obtained synchronization header as the starting point, a frame of spectral data (2048 bytes in total) is obtained, frame verification is performed on the obtained spectral data, and the position of the scientific payload data therein is obtained, and the spectral data and auxiliary information therein are extracted. The auxiliary information includes positioning and orbit determination data, payload compartment and platform compartment status data, payload working parameters, etc. The extracted spectral data has 6 segments in total, and the JPEG2000 algorithm is used to analyze the 6 segments of spectral data respectively, and the 6 segments of spectral images that have been analyzed are synthesized to obtain a complete spectral image;
[0045] Since nearly 2,000 data extraction processes need to operate on spectral data, it is necessary to create a data extraction process pool to manage the data extraction processes, and a buffer is required to accommodate the parsed spectral images. In an embodiment of the present invention, a producer-consumer model is adopted to manage the data extraction processes and buffers. The producer process is responsible for parsing the image, and the consumer process is responsible for reading and writing data to the disk.
[0046] It should be further explained that, in a specific implementation process, a corresponding image intensity sequence is obtained according to the analyzed spectral image, and a process of performing gradient calculation on the obtained image intensity sequence to obtain a corresponding image gradient sequence includes:
[0047] There are two commonly used scanning methods. One is to control the guide rail to move the imaging mirror to scan the sun. This scanning method controls the step length more accurately and can record the number of current scans and the number of steps in this scan. The other is to move the satellite along the direction of the sun without moving the slit and the imaging mirror to perform a comprehensive scan of the sun. This scanning method cannot record the number of scans and the current number of steps in this scan.
[0048] For the first method, when the scanning step number is 0, the scan starts from the top of the sun, and when the step number is n-1, the scan ends at the bottom of the sun. In the middle, a total of n spectral images are obtained. Then, the spectral image sequence {P0, P1, ..., P n-1} Stitching them together in sequence to form a full disk image. For the second method, since the scanning sequence count and scanning step count are not recorded, the scanning cycle can only be counted by the characteristics of the image to further determine the start and end positions of a scanning sequence. The present invention is aimed at the second method and is also applicable to other scanning methods that cannot perform sequence counting;
[0049] The image intensity of each frame of the obtained spectral image is calculated to obtain the corresponding image intensity sequence, which is recorded as {I0, I1, ..., I n-1};
[0050]
[0051] Where w is the width of the spectral image, h is the height of the spectral image, for The intensity value of the corresponding pixel at the position is calculated, and the gradient of the obtained image intensity sequence is calculated to obtain the corresponding image gradient sequence, which is recorded as {T0, T1, ..., T n-1};
[0052]
[0053] It should be further explained that, in a specific implementation process, the process of searching for the maximum image frame number in the obtained image gradient sequence, obtaining the scanning sequence according to the searched maximum image frame number, and then generating the full disk image includes:
[0054] In the obtained image gradient sequence, the image frame number corresponding to the position with the largest gradient change is marked as the maximum image frame number, that is, the starting and ending positions of each scan. The maximum image frame number includes two. During conventional observation, the scan starts from the north pole of the sun. At the beginning, it is a cold space, and then gradually enters the surface of the sun, and the surface becomes larger and larger, and then gradually decreases to the south pole, and then to the cold space again. When scanning to the center of the sun's sphere, the imaging pixels are the most, the average pixel value is the highest, and the two ends are the smallest;
[0055] Since there are different activities on the solar surface, the occurrence of activities will affect the average intensity, so the present invention adopts a gradient method. When on the solar surface, the gradient changes are not large, but when entering the solar surface from the cold space or entering the cold space from the solar surface, the gradient changes are relatively large. A scanning sequence is determined by these two large gradient changes. For a scanning sequence, it is spliced according to its internal wavelength order to obtain the corresponding full-disk image.
[0056] It should be further explained that, in the specific implementation process, the generated full disk image is traversed to obtain full disk pixel values, and the binary method is used to calculate the obtained full disk pixel values to obtain left and right extreme value points and upper and lower extreme value points. The process includes:
[0057] The obtained full disk image is subjected to gradient calculation. The place with the largest gradient change in each row or column of pixels is the edge of the sun. There are two gradient extreme value points when entering and leaving the sun, and the center coordinates of the two points are on the diameter of the sun. In the embodiment of the present invention, the possible center coordinates are determined by taking two points on the same straight line on the circumference of the full disk image. At this time, only row pixels or column pixels need to be obtained, and the first maximum gradient value and the last maximum gradient value inside the pixel are searched and recorded, so that the corresponding sun center position can be obtained. In this process, only row pixels or column pixels at the edge of the full disk image need to be obtained. Therefore, the binary search method is used when calculating the gradient to obtain the gradient extreme value points on both sides of the sun center without scanning the entire full disk image.
[0058] A traversal operation is performed on the obtained full disk image, and the traversal operation is used to visit each pixel point in the full disk image one by one to obtain the corresponding full disk pixel value. The obtained full disk pixel value is calculated by using a binary search method, and the number of pixel points in each row is divided equally, and the calculation is started from the starting and ending positions of each row to obtain the corresponding maximum gradient value. The maximum gradient value of each row includes two, and the two are marked as left and right extreme value points. The same method is used to divide the number of pixel points in each column equally, and the calculation is started from the starting and ending positions of each column to obtain the corresponding maximum gradient value. The maximum gradient value of each column includes two, and the two are marked as upper and lower extreme value points.
[0059] It should be further explained that, in the specific implementation process, the candidate heliocentric pixel positions are calculated based on the obtained left and right extreme points and the upper and lower extreme points, and the corresponding number of occurrences is recorded. The process of obtaining the heliocentric pixel coordinates based on the recorded number of occurrences includes:
[0060] Take any left and right extreme points as an example and mark them as (x L , y) and (x R , y), obtain the candidate heliocentric pixel position according to the coordinates of the left and right extreme points , adopt the same method to obtain the candidate heliocentric pixel positions of each left and right extreme point, and record the number of occurrences of each candidate heliocentric pixel position;
[0061] Take any upper and lower extreme points as an example and mark them as (x, y u ) and (x, y d ), obtain the candidate heliocentric pixel position according to the coordinates of the upper and lower extreme points , the same method is used to obtain the candidate heliocentric pixel positions of each upper and lower extreme point, and the number of occurrences of each candidate heliocentric pixel position is recorded. The candidate heliocentric pixel position with the largest number of occurrences is used as the heliocentric pixel position of the full solar disk image, and the corresponding heliocentric pixel coordinates are output.
[0062] It should be further explained that, in a specific implementation process, the process of obtaining the geometric distance between the heliocentric pixel coordinates and the image center coordinates and adjusting the generated full disk image according to the obtained geometric distance includes:
[0063] Obtain the image center of the full disk image and its corresponding image center coordinates, and mark the heliocentric pixel coordinates and image center coordinates as (x 日 ,y 日 ) and (x 图 ,y 图 ), the geometric distance between the heliocentric pixel coordinates and the image center coordinates is obtained, denoted as S;
[0064]
[0065] The orientation information of the heliocentric pixel coordinates relative to the image center coordinates is obtained, and the orientation information is used to describe the deviation angle of the heliocentric pixel coordinates relative to the image center coordinates. The obtained full solar disk image is moved by a corresponding geometric distance according to the deviation angle to achieve adjustment.
[0066] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.
Claims
1. A method for quickly parsing satellite downlink data, characterized in that: The following steps are involved: Step S1: receiving and decompressing the raw data to be processed to obtain decompressed data, segmenting the obtained decompressed data to obtain segmented data, extracting data from each segmented data to obtain spectral data, and analyzing and synthesizing the extracted spectral data to obtain a spectral image; Step S2: obtaining a corresponding image intensity sequence according to the analyzed spectral image, performing gradient calculation on the obtained image intensity sequence to obtain a corresponding image gradient sequence, searching for a maximum image frame number in the obtained image gradient sequence, obtaining a scanning sequence according to the searched maximum image frame number, and then generating a full disk image; Step S3: performing a traversal operation on the generated full disk image to obtain full disk pixel values, adopting a binary method to calculate the obtained full disk pixel values to obtain left and right extreme value points and upper and lower extreme value points, calculating candidate heliocentric pixel positions according to the obtained left and right extreme value points and upper and lower extreme value points, and recording the corresponding number of occurrences, obtaining heliocentric pixel coordinates according to the recorded number of occurrences, obtaining the geometric distance between the heliocentric pixel coordinates and the image center coordinates, and adjusting the generated full disk image according to the obtained geometric distance; The process of segmenting the obtained decompressed data to obtain segmented data includes: The original data transmitted by the satellite is received, and the Tar command provided by the Linux system is used to decompress the received original data to obtain corresponding decompressed data, and the obtained decompressed data is segmented, wherein the segmentation refers to dividing the decompressed data into a plurality of segmented data; The process of parsing and synthesizing the extracted spectral data to obtain a spectral image includes: A multi-process parallel processing method is used to extract data from each segmented data. The data extraction refers to extracting the spectral data from the original data. One segmented data corresponds to one data extraction process, and the position of the scientific payload data in the segmented data is obtained. The spectral data and auxiliary information therein are extracted. The JPEG2000 algorithm is used to analyze the six segments of spectral data respectively, and the six segments of spectral images that have been analyzed are synthesized to obtain a complete spectral image. Create a data extraction process pool to manage the data extraction process, create a buffer to hold the parsed spectral image, and use the producer-consumer model to manage the data extraction process and buffer. The producer process is responsible for parsing the image, and the consumer process is responsible for reading and writing data to the disk. The process of performing gradient calculation on the obtained image intensity sequence to obtain the corresponding image gradient sequence includes: The image intensity of each frame of the obtained spectral image is calculated to obtain the corresponding image intensity sequence, which is recorded as {I0, I1, ..., I n-1 }; ; Where w is the width of the spectral image, h is the height of the spectral image, for The intensity value of the corresponding pixel at the position is calculated, and the gradient of the obtained image intensity sequence is calculated to obtain the corresponding image gradient sequence, which is recorded as {T0, T1, ..., T n-1 }; 。 2. A method for rapid analysis of satellite downlink data according to claim 1, characterized in that: The process of searching for the maximum image frame number in the obtained image gradient sequence, obtaining a scanning sequence according to the found maximum image frame number, and then generating a full disk image includes: In the image gradient sequence, the image frame number corresponding to the position with the largest gradient change is taken as the maximum image frame number. The maximum image frame number includes two. A scanning sequence is determined according to the two maximum image frame numbers obtained. For a scanning sequence, it is spliced according to its internal wavelength order to obtain the corresponding full-disk image.
3. A method for rapid analysis of satellite downlink data according to claim 2, characterized in that: The generated full disk image is traversed to obtain full disk pixel values, and the process of using a binary method to calculate the obtained full disk pixel values to obtain left and right extreme value points and upper and lower extreme value points includes: A traversal operation is performed on the obtained full disk image, and the traversal operation is used to visit each pixel point in the full disk image one by one to obtain the corresponding full disk pixel value, and the obtained full disk pixel value is calculated by using the binary method, the number of pixels in each row is divided equally, and the calculation is started from the starting and ending positions of each row to obtain the corresponding left and right extreme value points, and the upper and lower extreme value points are obtained by the same method.
4. The method for quickly parsing satellite downlink data according to claim 3, characterized in that: The candidate heliocentric pixel positions are calculated based on the obtained left and right extreme points and upper and lower extreme points, and the corresponding number of occurrences is recorded. The process of obtaining the heliocentric pixel coordinates based on the recorded number of occurrences includes: The coordinates of the left and right extreme points of each group are obtained, and the candidate heliocentric pixel positions are obtained according to the coordinates of the left and right extreme points, and the number of occurrences of each candidate heliocentric pixel position is recorded; the coordinates of the upper and lower extreme points of each group are obtained, and the candidate heliocentric pixel positions are obtained according to the coordinates of the upper and lower extreme points, and the number of occurrences of each candidate heliocentric pixel position is recorded; the candidate heliocentric pixel position with the largest number of occurrences is used as the heliocentric pixel position of the full-sun disk image, and the corresponding heliocentric pixel coordinates are output.
5. A method for rapid analysis of satellite downlink data according to claim 4, characterized in that: The process of obtaining the geometric distance between the heliocentric pixel coordinates and the image center coordinates and adjusting the generated full disk image according to the obtained geometric distance includes: The image center of the full-disk image and its corresponding image center coordinates are obtained, the geometric distance between the two is obtained according to the obtained heliocentric pixel coordinates and the image center coordinates, the azimuth information of the heliocentric pixel coordinates relative to the image center coordinates is obtained, the azimuth information is used to describe the deviation angle of the heliocentric pixel coordinates relative to the image center coordinates, and the obtained full-disk image is moved by a corresponding geometric distance according to the deviation angle to achieve adjustment.
Citation Information
Patent Citations
Solar activity recognition method based on scale transformation model
CN106570506A
a solar image registration method based on normalized cross-correlation and SIFT
CN109785371A
